§ case
Lit Law
Jul 2023 — Jul 2025
AI · Legal-Tech · Full Stack · Backend Lead
- —A production AI-powered contract-drafting platform for Indian law — backend architected as Node.js/Express microservices over gRPC, with PostgreSQL (Drizzle ORM) and Redis caching behind 40+ JWT-authenticated REST APIs.
- —RAG pipelines on custom embeddings via Pinecone vector search, cutting retrieval latency 48% at 1,500+ concurrent users.
- —Real-time collaborative editing — presence and concurrent edits — delivered over SSE to a Next.js (SSR) frontend built with React, Redux, Tailwind, and Tiptap.
- —Payments integrated across two providers (Razorpay, Cashfree) at 100% reliability across 500+ signings/month, backed by 95%+ test coverage.
- —Deployed as Docker containers on GCP instances.
Node.jsExpress.jsPostgreSQLRedisDrizzle ORMgRPCPineconeNext.jsReactReduxTailwind CSSTiptapRazorpaySSEDockerGoogle Cloud Platform (GCP)
◼ Challenges
- →Latency, throughput, and uptime targets on a production legal-AI platform
- →Real-time concurrent editing without server-state nightmares
- →Reliable payments at scale across two providers
◼ Solutions
- →Microservices over gRPC with PostgreSQL + Drizzle; 40+ JWT REST APIs and 95%+ test coverage
- →RAG with custom embeddings on Pinecone, 48% latency reduction at 1,500+ concurrent users
- →SSE-based collaboration; Razorpay + Cashfree integrated at 100% reliability across 500+ signings/mo